ncoplin_at_orbeng.com
Date: 2002-10-31 05:15:44
Hi Cameron,
I was just catching up with old postings and this idea caught my attention.
C't Aug1985 (German magazine) had an article about simulation of a 65C02
with a software 6502 emulator, and so it raises the question how about a
software 65816. With an REU some of the block-move type functions would
execute very fast, and the 16bit indexing could be done with zp
instructions.... as you say with a virtual memory manager you could use any
RAM expansion to pretend to be real RAM.
I did have some questions about the actually code you used:
1) what did your decode routine look like. How did decode the op code and
jmp to the right routine (an efficient way of doing this would be handy for
some other coding projects...)
2) what do you estimate the emulated through-put was relative to the 1MHz
base, ie did your sample program run effective at 0.2Mhz or slower/faster?
Regards, Nick
-----Original Message-----
From: Cameron Kaiser [mailto:spectre@stockholm.ptloma.edu]
Sent: Wednesday, 2 October 2002 1:04 AM
To: cbm-hackers@cling.gu.se
Subject: Software 6502-on-6502
Well, last night was a very productive all-night hack session. (Medical
students aren't *always* studying.)
In between cranking more speed out of Nether G2's core renderer (10fps!
wahoo!), I finally threw together some ideas that had been floating through
my head about a software 6502-on-6502 emulator (i.e., an emulator in 6502
assembly that emulates a 6502). While this might seem like an academic
pursuit at best, I think it would be the most elegant way of adding things
to the 6502 that we don't have yet such as true MMU facilities, protected
memory, virtual memory, etc.
In any case, it was functional enough to run this:
testp lda $d020
sta $d021
tay
iny
sty $02
ldx #0
lda #1
tlup sta $0400,x
sta $0500,x
sta $0600,x
sta $0700,x
inx
bne tlup
brk
All memory access is trapped and checked, all instructions are byte-code
verified, and as you can see, immediate, absolute, implied, branch and
absolute,x addressing modes are all working. I just need to finish the rest
of the opcodes, which were pretty easy (just distill an instruction into a
safe variant, and then run that variant, which sets all the proper flags and
whatnot for me), as well as the other addressing modes.
While I was hand-typing the instruction decode table, there were some
intriguing instruction patterns that came out. The table is below. I don't
use any of these generalisations for decoding (just a simple raw lookup),
but they might be useful to someone trying to write a better (dis)assembler.
; master instruction decode table
vectab .word opbrk, opora, opbad, opbad
.word opbad, opora, opasl, opbad ; $00-7
.word opphp, opora, opasl, opbad
.word opbad, opora, opasl, opbad ; $08-f
.word opbpl, opora, opbad, opbad
.word opbad, opora, opasl, opbad ; $10-7
.word opclc, opora, opbad, opbad
.word opbad, opora, opasl, opbad ; $18-f
.word opjsr, opand, opbad, opbad
.word opbit, opand, oprol, opbad ; $20-7
.word opplp, opand, oprol, opbad
.word opbit, opand, oprol, opbad ; $28-f
.word opbmi, opand, opbad, opbad
.word opbad, opand, oprol, opbad ; $30-7
.word opsec, opand, opbad, opbad
.word opbad, opand, oprol, opbad ; $38-f
.word oprti, opeor, opbad, opbad
.word opbad, opeor, oplsr, opbad ; $40-7
.word oppha, opeor, oplsr, opbad
.word opjmp, opeor, oplsr, opbad ; $48-f
.word opbvc, opeor, opbad, opbad
.word opbad, opeor, oplsr, opbad ; $50-7
.word opcli, opeor, opbad, opbad
.word opbad, opeor, oplsr, opbad ; $58-f
.word oprts, opadc, opbad, opbad
.word opbad, opadc, opror, opbad ; $60-7
.word oppla, opadc, opror, opbad
.word opjmp, opadc, opror, opbad ; $68-f
.word opbvs, opadc, opbad, opbad
.word opbad, opadc, opror, opbad ; $70-7
.word opsei, opadc, opbad, opbad
.word opbad, opadc, opror, opbad ; $78-f
.word opbad, opsta, opbad, opbad
.word opsty, opsta, opstx, opbad ; $80-7
.word opdey, opbad, optxa, opbad
.word opsty, opsta, opstx, opbad ; $88-f
.word opbcc, opsta, opbad, opbad
.word opsty, opsta, opstx, opbad ; $90-7
.word optya, opsta, optxs, opbad
.word opbad, opsta, opbad, opbad ; $98-f
.word opldy, oplda, opldx, opbad
.word opldy, oplda, opldx, opbad ; $a0-7
.word optay, oplda, optax, opbad
.word opldy, oplda, opldx, opbad ; $a8-f
.word opbcs, oplda, opbad, opbad
.word opldy, oplda, opldx, opbad ; $b0-7
.word opclv, oplda, optsx, opbad
.word opldy, oplda, opldx, opbad ; $b8-f
.word opcpy, opcmp, opbad, opbad
.word opcpy, opcmp, opdec, opbad ; $c0-7
.word opiny, opcmp, opdex, opbad
.word opcpy, opcmp, opdec, opbad ; $c8-f
.word opbne, opcmp, opbad, opbad
.word opbad, opcmp, opdec, opbad ; $d0-7
.word opcld, opcmp, opbad, opbad
.word opbad, opcmp, opdec, opbad ; $d8-f
.word opcpx, opsbc, opbad, opbad
.word opcpx, opsbc, opinc, opbad ; $e0-7
.word opinx, opsbc, opnop, opbad
.word opcpx, opsbc, opinc, opbad ; $e8-f
.word opbeq, opsbc, opbad, opbad
.word opbad, opsbc, opinc, opbad ; $f0-7
.word opsed, opsbc, opbad, opbad
.word opbad, opsbc, opinc, opbad ; $f8-f
--
----------------------------- personal page: http://www.armory.com/~spectre/
--
Cameron Kaiser, Point Loma Nazarene University *
ckaiser@stockholm.ptloma.edu
-- Math according to Pentium: 2 / 2 = 1.037587439439485486372112039523781385
...
Message was sent through the cbm-hackers mailing list
PLEASE TAKE NOTE:
The contents of this email (including any attachments) may be
privileged and confidential. Any unauthorised use of the contents
is expressly prohibited. If you have received this email in error,
please advise us immediately (you can contact us by telephone
on +61 8 9441 2311 by reverse charge) and then permanently
delete this email together with any attachments. We appreciate
your co-operation.
Whilst Orbital endeavours to take reasonable care to ensure
that this email and any attachments are free from viruses or other
defects, Orbital does not represent or warrant that such is explicitly
the case
(C) 2000: Orbital Engine Company (Australia) PTY LTD and its
affiliates
Message was sent through the cbm-hackers mailing list
Archive generated by hypermail 2.1.4.